In their essay Longtermism in an Infinite World (2025), Christian Tarsney and Hayden Wilkinson address a significant challenge to longtermist ethics, exploring decision-making in a possible future containing an infinite number of value locations (token entities of some common type - persons & non-persons - that can exist across different outcomes, with which valuable events are associated). The authors propose a formal framework in which moral evaluation can be extended to infinite temporal domains, arguing that agents can still act meaningfully by influencing ‘finite’ parts of an otherwise infinite world. This claim aims to reconcile longtermism’s commitment to the moral significance of a distant future with the difficulties posed by infinite value aggregation. The framework posed in the chapter thus presupposes a metaphysical conception of time which treats all temporal locations as equally real and comparable within a single evaluative structure, aligning closely with eternalism. The eternalist view posits that past, present, and future, are ontologically equivalent and thus exist in a static four-dimensional spacetime block, generating a tension at the very heart of longtermist reasoning: assuming that the future exists in the same way as the present does, how can moral agents meaningfully affect it? To get around this core issue, Tarsney and Wilkinson appeal to ‘affectable’ parts of the universe, relying on an implicit causal asymmetry that does not comply with the basis of an eternalist ontology, upon which longtermism is based.
This essay will argue that the coherence of the Infinite World framework discussed in chapter 6 of Essays on Longtermism: Present Action for the Distant Future (Greaves et al., 2025), relies on a metaphysical ambiguity between eternalism and more dynamic conceptions of time. By creating a dialogue between the metaphysic of time and longtermist ethics, I aim to suggest that the apparent paradoxes of infinite value and infinite agency arise not merely from mathematical formalism, but from underlying temporal assumptions.
Longtermism is the thesis that the moral value of present actions is primarily determined by their potential effect on the far future, and thus that we should prioritise actions that most improve the long-term future. This aspect is explored by Tarsney and Wilkinson in Longtermism in an Infinite World (2025), throughout which they investigate whether longtermism can remain coherent and action-guiding if the universe contains an infinite amount of morally relevant value locations. They argue that totals moral theories (which sum value across all locations) face severe problems in an infinite universe as comparisons of ‘better’ or ‘worse’ outcomes become undefined if every option holds infinite value, claiming that risk neutral totals (the view that we must maximise the expected total value regardless of variance) breaks down under such conditions. As a solution to this crucial issue, the authors argue that moral evaluation should not depend on the entire universe, as this may be infinite and beyond our control, but solely on the finite portion that our actions can actually influence. They hence divide the universe into A (the finite affectable part: locations whose values depend on our action, such as the next billion years), and U (the infinite unaffectable part: locations whose values are the same no matter what actions we commit, namely, distant galaxies outside of our causal reach). In short, the chapter argues that longtermism does not require comparing total infinities, as this would cause an undefined result when placed in a mathematical equation, but should solely compare marginal effects within the causal reach of our actions. This restriction hence allows moral comparisons in the case of an infinite universe through the conclusion that while the universe as a whole may have infinite value, the difference our choices make are finite and measurable, preserving the practical action-guidance of longtermism, whilst avoiding the mathematical indeterminacy of infinite sums (which I will not go into, considering my misalignment with mathematical knowledge…).
The model discussed by Tarsney and Wilkinson is based on what is called Risk Neutral Totalism (RNT) and requires an axiology which is impartial, meaning that all value locations receive the same weight in total sum of value regardless of their spatiotemporal location. This premise is essential given that longtermism rests on the vast number of these value locations in the far future. This commitment to temporal impartiality implies that future value is equally morally significant to present value, hence referring to a moral analogue of Eternalism (where past, present and future are ontologically symmetric). The authors highlight that the rejection of RNT would lead to alternative axiologies that are less favourable to longtermism. An example brought of this is Pure Time Discounting, which abandons RNT’s commitment to impartiality by reducing value based on a location’s position in time. By rejecting this theory, Tarsney and Wilkinson affirm the eternalist-friendly scope of their axiology. By claiming the future to be as real as the present in an attempt to salvage longtermism from the infinite ethics problem, causal influence seems to be a mere internal relation within the block of the existing universe rather than a changeable effect. The affectable part of the future, central to Tarsney & Wilkinson’s argument, becomes metaphysically puzzling if nothing is truly affectable, given that the block presumed by the model is fixed. What we believe to improve through our actions, is only one region of the block which is correlated with another (future) region, instead of being a process of change or influence. This conclusion implicates that the authors’ use of finite predictable effects implicitly depends on A-theoretic causation, inherently absent in Eternalism, but found in more dynamic temporal ontologies.
To conclude, Tarsney and Wilkinson’s attempt to reconcile longtermism with the possibility of an infinite universe succeeds in preserving mathematical coherence but does so at a significant metaphysical cost. While this framework depends on an implicit eternalist ontology, it undermines the causal asymmetry that is required for moral agency and by the practical guidelines of the longtermist view. The ‘affectable’ parts of the infinite world hence appear as fixed relations within what is an unchanging and equal temporal block, rather than as domains of influence. By observing the metaphysics of time, we see that the challenges facing longtermism are not merely mathematical (infinite sums) but ontological, explicating that the nature of the future determines whether it can be morally shaped. This incoherence in the chapter hence suggests the need of a more dynamic conception of time as a foundation for longtermist ethics, allowing us to question whether it is better to inhabit a future where our present actions are constrained to finite effects, or one where our physical reality forces us to face the mathematical difficulties of infinite axiology.
